Secretome-Mediated Neuroprotective Effects of the hiPSC-derived Glial and Neuronal Progenitor Cells in the Hypoxia-Induced Neuronal Damage (Comparative Study)

Background: Stem cell secretomes hold great promise for regenerative medicine. This study is focused on the secretome-mediated neuroprotective effects of the human induced pluripotent stem cell-derived neuronal and glial progenitor cells. Therapeutic properties of the secretomes were assessed under conditions of the hypoxia-induced neuronal damage in vitro and in vivo. Methods: Secretory activity of the cultured neuronal and glial progenitor cells was analyzed by proteomic and immunosorbent-based approaches. Conditioned media collected from the cultures was tested for neuroprotective properties in vitro and in vivo. In vitro experiments involved exposure of SH-SY5Y cells to the conditioned media during the recovery from the cobalt chloride-induced hypoxia. Neuroprotective effects were assessed by cell survival and neurite outgrowth. Cell survival indicators included MTT and LDH tests, vital staining with propidium iodide and Hoechst 33342, and polymerase chain reaction assay for the expression of apoptosis-related genes. Neurite outgrowth was assessed by alterations in SH-SY5Y cell morphology and MAP2/GAP43 gene expression dynamics. In vivo experiments involved intra-arterial administration of the conditioned media to laboratory rats during the recovery from experimental ischemic stroke. Neuroprotective effects were assessed by overall survival, neurologic decit and infarct volume dynamics, as well as by the end-point values of the apoptosis- and inammation-related gene expression levels, the extent of microglia/macrophage inltration, and the numbers of newly formed blood vessels in the affected area of the brain. Results: Secretomes of glial and neuronal progenitor cells partially overlapped, with specic proteins (found in secretome of one of the studied cultures and absent from the other) constituting, respectively, 31% and 45%. The glial progenitor cell-conditioned media showed higher content of neurotrophins (BDNF, GDNF, CNTF and NGF).


Introduction
Deterioration of brain function in ischemic stroke is overwhelming. The acute hypoxic damage to brain tissue causes focal and generalized neuronal death with diverse neurological sequelae [1]. Effectiveness of treatment and rehabilitation of the patients after acute cerebrovascular episodes remains extremely low. In the absence of etiological treatment, it is important to promote rehabilitation of the patients by stimulating nervous tissue repair [2]. New e cient maintenance therapies for cerebrovascular diseases are highly relevant.
Stem cells represent a promising and rapidly developing tool of regenerative medicine. The e ciency of cell therapies is based on two fundamental properties of the transplant -its ability to replace the destroyed differentiated cells of the tissue (replacement mechanism) and its ability to secrete regulatory molecules for the processes of in ammation, angiogenesis and regeneration (paracrine mechanism) [3,4].
The paracrine concept of cell therapies is currently taking the lead [5,6]. Accordingly, the conditioned tissue culture media become of increasing interest as a source of biologically active molecules.
Neurotrophins, cytokines and other regulatory molecules which accumulate in the tissue culture media can exert complex positive effects in various pathological conditions [4]. Good prospects for the use of stem cell-conditioned media for the treatment of neurodegenerative and vascular diseases was demonstrated in a number of studies [7][8][9].
Neuronal and glial progenitor cells (respectively, NPCs and GPCs) can be derived from human induced pluripotent stem cells (hiPSCs). Positive in uence of their secretomes on the damaged nervous tissue appears plausible. This in uence has not yet been studied and no substantive information on this subject can be found in the literature.
Here we report the results of comparative assessment of neuroprotective paracrine effects of the hiPSCderived NPCs and GPCs. Therapeutic relevance of the NPC-and GPC-conditioned media was assessed in vitro and in vivo by using, respectively, SH-SY5Y model of cobalt-induced hypoxia and rat experimental model of ischemic stroke.
To obtain the conditioned media (CM), NPС and GPС cultures were washed twice with PBS and incubated in DMEM/F12 (PanEco) for 12 h. The medium was collected and centrifuged for 5 min at 3000 rpm. The supernatant was passed through 0.2 µm syringe lter (PanEco) and used in the experiments.

Flow cytometry
The cells were detached with Versene solution (PanEco) and pelleted by centrifugation at 1800 rpm for 5 min, then washed with HBSS (PanEco), xed in 4% paraformaldehyde for 10 min, washed with PBS, permeabilized with 70% methanol on ice for 10 min, washed twice with PBS and collected by centrifugation at 1800 rpm for 5 min. Fixed cells were incubated with primary antibodies to PAX6 (ab5790, Abcam), glial marker S100b (ab52642, Abcam) or neuronal marker βIII tubulin (ab182070, Abcam) at + 4 °С for 12 h, then washed with PBS, collected by centrifugation at 1800 rpm for 5 min and incubated with secondary antibodies (Alexa Fluor 488 conjugated, Invitrogen) for 60 min in the dark. Fixed cells exposed to secondary antibodies only were used as a negative control for the ow cytometrybased quanti cation. Stained cells were analyzed on a CyFlow ML ow cytometer with FloMax software (Partec, Germany).

Proteomic analysis
Ultimate 3000 RSLCnano HPLC system (Thermo Scienti c, USA) connected to Q-exactive HF mass spectrometer (Thermo Scienti c) was used for the analysis. Mass spectra were recorded in the ionpositive mode with nanoelectrospray ionization. Identi cation of proteins in the spectra was carried out in SearchGUI v.3.3.16 software with X!Tandem, OMSSA and MS-GF + algorithms (Vaudel M. et al., 2011). The alignments were made with Uniprot database ('human'-ltered). Structuring of the data by classes and functions of the proteins was carried out by using Uniprot and PANTHER databases [15].
Enzyme-linked immunosorbent assay (ELISA) Concentrations of GDNF, BDNF, NGF and CNTF in CM were measured by ELISA (RnD systems, USA) in accordance with the manufacturer's protocol. The collected CM samples were concentrated 24-fold with 3 kDa Amicon Ultra lter units (Sigma-Aldrich) to 0.5 mg/mL total protein concentrations prior to the analysis. Optical densities (absorbance at 450 nm) were measured in a plate reader (PerkinElmer, USA).

PCR assay
For the reverse transcription polymerase chain reaction (RT-PCR) assay, cells or tissues were collected in RNAlater (Thermo Fisher Scienti c, USA) for the preservation of RNA molecules during storage. Total RNA from the collected cell or tissue samples was isolated with RNeasy Mini Kit (Qiagen, USA) according to the manufacturer's protocols. cDNA synthesis was carried out with RevertAid First Strand cDNA Synthesis Kit (Thermo Fisher Scienti c) according to the manufacturer's protocol. The real-time PCR mixtures were set up with qPCRmix-HS SYBR (Evrogen, Russia); the reactions were carried out in a Bio-Rad iQ thermal cycler as primary denaturation 95 °C for 5 min followed by 40 cycles of denaturation 95 °C for 20 seconds, annealing 55-63 °C for 20 seconds and elongation 72 °C for 20 seconds. Raw data for the genes of interest were normalized against constitutively expressed reference genes GAPDH (Gapdh, glyceraldehyde-3-phosphate dehydrogenase) and ACTB (Actb, beta-actin). Expression levels were calculated by 2 −ΔΔC(T) approach. The oligonucleotide sequences are given in Supplementary Table 1.

In vitro and in vivo study design
In vitro experiments involved SHSY-5Y neuroblastoma cell line. CM were applied at 4 h after hypoxia modeling to the nal total protein concentration of 5 µg/mL. At 24 h after hypoxia modeling, cell survival was assessed by MTT and LDH tests, vital staining with propidium iodide and Hoechst 33342, and the levels of apoptosis-related gene expression. In the longer incubation experiments, CM were added every other day to the same nal protein concentration. Formation of axons and dendrites was assessed by expression markers of neuritogenesis MAP2 and GAP43 and alterations in cell morphology on days 1, 3 and 7 of the experiment.
In vivo experiments involved adult male Wistar rats of 250-300 g body weight (n = 34) purchased from AlCondi Ltd., Russia. All manipulations with the animals were approved by the Ethical Committee at N.I. Pirogov Russian National Research Medical University. All in vivo experiments were carried out in accordance with EU Directive 2010/63/EU. The experimental ischemic stroke was modeled by temporary occlusion of the right middle cerebral artery. At 24 h after the stroke, the animals (n = 34) were randomized into 3 groups: control group (n = 10) and CM recipients (n = 12 each) to receive intra-arterial infusions of, respectively, the non-conditioned DMEM/F12-based medium, concentrated GPC-CM or NPC-CM to 50 µg/mL of total protein. Therapeutic effects were assessed the following parameters: survival, neurological de cit and the stroke volume immediately before the infusion (day 1) and on days 7, 14 and 30 post infusion (p/i). The animals were sacri ced by overdose of diethyl ether at the end of the experiment (day 30 p/i). At the autopsy, the ischemized portions of the brain tissue were dissected into 3 fragments. One fragment was xed in 10% formalin for para n sections, one was preserved in liquid nitrogen for cryosections, and one was preserved in RNAlater reagent for PCR analysis. Corresponding area from the unaffected contralateral hemisphere was dissected in the same way to be used as a control.
Evaluation of cell survival MTT ((3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide, PanEco) was added to the culture medium to a nal concentration of 0.5 mg/mL. The measurements were carried out after 2 h incubation with MTT and subsequent dissolution of formazan crystals with DMSO (PanEco). Evaluation of necrosis rates by LDH tests was carried out by using LDH Activity Assay Kit (Sigma-Aldrich) according to the manufacturer's protocol. Optical densities (absorbance at 570/620 nm for MTT test or 450 nm for LDH test) were measured in a plate reader (PerkinElmer). Cell death rates were assessed by staining the cells with vital uorescent dyes Hoechst 33342 (Sigma-Aldrich, 5 µg/mL for 30 min) and propidium iodide (Sigma-Aldrich, 1 µg/mL for 10 min) to visualize nuclear fragmentation and membrane integrity. The uorescence patterns were observed and recorded with Axio Observer.D1 inverted uorescence microscope equipped with AxioCam HRc (Carl Zeiss).

Assessment of neurite outgrowth
Formation of axons and dendrites was assessed by morphological examination combined with immunocytochemistry for βIII tubulin, and also by real-time PCR assay for the transcriptomic markers of neuritogenesis (MAP2, GAP43).

In vivo study of neuroprotective properties
Modeling of ischemic stroke Transient occlusion of the middle cerebral artery in rats was performed by standard method (the procedure originally developed by Koizumi [17] and modi ed by Longa [18]). The animals were anesthesized with 2.5-3% iso urane/97-97.5% air mixture. The right middle cerebral artery was temporary occluded with a silicone rubber-coated mono lament (diameter 0.19 mm, length 30 mm, diameter with coating 0.37 ± 0.02 mm, coating length 3-4 mm; Doccol, USA) for 90 min. During the operation and until the emergence from anesthesia the body temperature was maintained at 37 °С with a warming pad.
Comparative assessment of the neurological de cit and functional recovery Therapeutic effects of CM administration were determined by survival rates, infarct volume and neurological de cits as assessed by using the modi ed neurological severity scores (mNSS) for rodents [19] immediately before the infusion (day 1) and on days 7, 14 and 30 p/i.

Magnetic resonance imaging (MRI)
The imaging was accomplished in a 7T ClinScan small animal MRI system (Bruker BioSpin, USA). Intraoperative MRI (monitoring of the mono lament position, intracranial blood ow and hemorrhagic complications) was carried out as described previously [20]. For MRI evaluation of in vivo dynamics of the stroke volume, axial plane T2-weighted brain images (Т2-WI) were acquired immediately before the infusion (day 0) and on days 7

Histology
The formalin-xed tissues were dehydrated and embedded in para n and cut on a microtome. The 5-7 µm thick sections were positioned on gelatin-coated slides and dried at 37 °С for 24 h. The sections were depara nized in xylene and rehydrated in 100°-70° ethanol series, stained with H&E, dehydrated and mounted. Pathohistological examination of the sections by light microscopy was carried out using Axio Observer.D1 equipped with AxioCam HRc (Carl Zeiss). Formation of new blood vessels at the site of injury was assessed in the images of the H&E stained histological sections at х400 magni cation by using ImageJ software. Blood vessels were identi ed as hollow structures of characteristic morphology and red blood cells in the limina. The total number of blood vessels per sq mm of the section and the volume density of blood vessels (estimated as the total cross-sectional area of blood vessels per sq mm of the section) were evaluated.

Immunohistochemistry
The tissues cryopreserved in liquid nitrogen were cryosectioned at a 4-5 µm thickness. The sections were pre-blocked in PBS with 0.3% Triton X-100 and 2% BSA for 1 h, incubated with anti-CD68 primary antibodies (ab125212, Abcam) at + 4 °C overnight, washed and incubated with secondary antibodies (Alexa Fluor 488 conjugated, Invitrogen) for 60 min in the dark. The nuclei were counterstained with DAPI solution (1 µg/mL in PBS). The images were recorded with Axio Observer.D1 inverted uorescence microscope equipped with AxioCam HRc (Carl Zeiss). Total counts of CD68 + cells per sq mm of the section were determined with ImageJ software by counting the positively stained cells at x200 magni cation.

Statistical analysis
The data were processed in SigmaPlot 12.5 software. Each single group included 3-5 independent experiments. Pairwise comparisons were carried out by t-test in the case of normal distribution or Mann Whitney test in the case of non-normal distribution. For multiple comparisons, one-way ANOVA with Holm-Sidak method was used for the cases of normal distribution, whereas ANOVA on ranks with Dunn's test was used for the cases of non-normal distribution. Differences at p ≤ 0.05 were considered signi cant.

Characterization of hiPSCs, NSCs, GPCs and NPCs
The obtained hiPSCs were morphologically similar to human embryonic stem cells with high nuclearcytoplasmic ratio (Fig. 1А). The cells expressed pluripotency markers OCT4, NANOG and SOX2, and were immunopositive for OCT4 and NANOG transcription factors and SSEA4 and TRA-1-81 proteoglycans. The cells also showed the capacity of spontaneous differentiation into derivatives of any of the three germ layers (ectoderm, endoderm and mesoderm), which con rmed their pluripotency at the functional level (Fig. 1A, B). Culturing with appropriate inducers (see Sect. 2.1) afforded neural stem cells (NSCs) -small, densely growing cells prone to the formation of 3D rosette-like structures. The NSC phenotype was con rmed immunocytochemically and by PCR assay. NSCs expressed molecular markers of neural differentiation (PAX6, SOX2 and NES) and were immunopositive for the corresponding proteins.
Differentiation e ciency (as the proportion of PAX6-positive cells evaluated by ow cytometry) constituted 98 ± 1.8% (Fig. 1C). Upon stimulation with glial inducers the cells acquired spindle-shaped morphology with uneven outlines and large oval nuclei. Stimulation of NSCs with neuronal inducers promoted the outgrowth of neurites up to three cell diameters long. Glial progenitor cells (GPCs) expressed S100B and GFAP, and 97 ± 2.8% of them were S100B-positive. Neuronal progenitor cells (NPCs) had smaller nuclei, expressed neuronal markers TUBB3, MAP2 and ENO2, and 96 ± 3.4% of them were βIII tubulin positive (Fig. 1D).

Evaluation of neuroprotective effects of NPC-CM and GPC-CM in the in vitro model
Modeling of hypoxia in SH-SY5Y cells by the 4 h exposure to 250 µM CoCl 2 dramatically affected their survival as indicated by МТТ tests carried out at 24 h after the treatment. In the absence of NPC-CM or GPC-CM, hypoxia caused a reduction in the numbers of viable cells by 73 ± 4.1% and an increase in the amounts of LDH released from necrotic cells by 11.7 ± 2.4% as compared with the no-hypoxia controls.
The exposure to NPC-CM increased the survival of SH-SY5Y cells by 9.26 ± 2.8%; however, the amounts of cell-free LDH measured after the exposure to NPC-CM and after hypoxia were similar. By contrast, the exposure to GPC-CM not only increased the SH-SY5Y cell survival by 25.4 ± 6.1% but also promoted a reduction in the levels of cell-free LDH in the culture medium by 8.7 ± 2.7% compared with the CoCl 2 treated cultures (Fig. 3A).
In the aftermath of CoCl 2 -induced hypoxia, the cells died predominantly by apoptosis (55.4 ± 3.9%) and to a lesser extent by necrosis (16.6 ± 2.5%, Fig. 3A). Treatment of the cells with GPC-CM reduced the apoptotic cell numbers to 35.1 ± 2.8% and necrotic cell numbers to 10.2 ± 3.9%. Treatment of the cells with NPC-CM reduced the apoptotic cell numbers to 42.2 ± 3.0% only and caused no signi cant reduction in the necrotic cell numbers.
Cell viability was additionally assessed by expression of apoptosis-related genes -pro-apoptotic BAX and anti-apoptotic BCL2, which encode regulatory proteins of the same family. At 24 h after the CoCl 2induced hypoxia, the level of BCL2 expression was reduced 6.2-fold whereas the level of BAX expression was increased 2.3-fold as compared with the no-hypoxia control. The treatment with GPC-CM promoted a 2-fold reduction in BAX expression and a 3.7-fold increase in BCL2 expression as compared with the CoCl 2 treated cultures. Interestingly, in terms of BAX/BCL2 expression, the effect of treatment with NPC-CM was similar: it promoted a 2-fold reduction in BAX expression and a 3.2-fold increase in BCL2 expression as compared with the CoCl 2 treated cultures (Fig. 3A).
In the no-hypoxia controls, the cells retained typical SH-SY5Y morphology with long βIII tubulinimmunopositive outgrowths. The CoCl 2 -induced hypoxia caused degeneration of axons and dendrites: neuronal bodies were prominent, but the neurites were missing or shortened (Fig. 3B).
In the aftermath of CoCl 2 -induced hypoxia without CM treatment, expression levels of MAP2 and GAP43 were signi cantly reduced during the entire period of observation (days 1, 3 and 7) as compared with the no-hypoxia control (Fig. 4C). On days 1 and 3, the NPC-CM treated cell cultures showed no signs of neurite outgrowth. By day 7, however, the NPC-CM treated cells developed short processes while having signi cantly elevated MAP2 and GAP43 expression levels as compared with the CoCl 2 treated cultures.
Neurotrophic effects of GPC-CM were more pronounced: regeneration of axons and dendrites in the GPC-CM treated cultures was evident during the entire period of observation (days 1, 3 and 7). On day 7, MAP2 and GAP43 expression levels in the GPC-CM treated cells were signi cantly elevated as compared with both the NPC-CM and CoCl 2 treated cultures (Fig. 4C).

Therapeutic effects of the intra-arterial infusion of NPC-CM and GPC-CM in the experimental ischemic stroke
Kaplan-Meier survival curves for three groups (NPC-CM treated, GPC-CM treated and the non-conditioned media-treated control) are shown in Fig. 4A. All deaths occurred within 3 days after the stroke and were associated with vasogenic cerebral edema. CM infusions had no effect on survival.
Neurological severity scores (mNSS) were recorded at the following time points: immediately before the infusion (day 1), and on days 7, 14 and 30 p/i. Neurological de cit reached its maximum by 24 h after the acute focal ischemia modeling in all groups; its progression was associated with the cerebral infarct development. After that, the neurological de cit underwent regression in all groups; the highest rates of functional recovery were observed during the initial 2 weeks of the experiment. The scores for the GPC-CM group on days 14 and 30 p/i were, respectively, 1.5-times and 1.6-times lower as compared with the control group; these differences were signi cant (Fig. 4B). By contrast, no signi cant differences in mNSS between the NPC-CM group and the control group were observed during the experiment. The data indicate enhanced functional recovery of brain function in response to the infusion of GPC-CM during the acute period of ischemic stroke.
The stroke volume was evaluated by MRI, with Т2-weighted brain magnetic resonance images acquired at different time points used for the evaluation. Reduction in the infarct volume was pronounced in all groups; no signi cant differences between the groups were revealed in the course of observation (Fig. 4C,  D).
To understand the CM-mediated therapeutic effects at molecular level, expression of apoptosis-and in ammation-related molecular markers in brain tissues was studied by PCR-based assay (Fig. 5). The GPC-CM group showed signi cantly reduced expression of the pro-apoptotic gene Bax compared with the control group. By contrast, expression levels of Bax in the ischemized tissue of the NPC-CM group were 2.4-times higher as compared with the intact brain tissue in contralateral cerebral hemisphere (IH) and the GPC-CM group; these differences were also signi cant. At the same time, no signi cant differences in Bcl2 expression were observed between the groups (Fig. 5A).
Vascular necrosis causes secondary damage to brain parenchyma due to the continuous in ammatory reaction accompanied by the elevated expression of Tnf gene. Infusion of GPC-CM speci cally reduced Tnf expression in the affected brain parenchyma; the difference with the control was statistically signi cant. In addition, expression levels of the anti-in ammatory cytokine genes Il4, Il10 and Il13 were signi cantly elevated in the ischemized brain tissues of GPC-CM treated animals compared with the control group. By contrast, expression levels of Tnf, Il4, Il10 and Il13 in the ischemized brain tissues of NPC-CM treated animals were comparable with the control group (Fig. 5A). Expression levels of Il1b and Il6 in all groups were similar.
As demonstrated by histological study, the infusion of GPC-CM supported neoangiogenesis. The counts of newly formed perfused blood vessels in the ischemized brain tissues of GPC-CM treated animals were the highest; the difference with the control group was signi cant (p < 0.05, Fig. 5B, C).
CD68 is a cell surface glycoprotein highly expressed on phagocytic cells of the resident microglia and in ltrating monocytic macrophages. Accumulation of CD68 positive cells in the ischemized brain area was detected at the autopsy on day 30 p/i in all groups. However, this effect was signi cantly alleviated in GPC-CM treated animals, as the numbers of accumulated phagocytic cells were lower compared with other groups. No such alleviation was observed in the NPC-CM group where the counts of СD68 + cells in the damaged area were signi cantly higher (Fig. 5B, C).

Discussion
hiPSC-derived NPCs and GPCs produce and secrete numerous regulatory proteins and peptides. NPC and GPC secretomes partially overlap, but large proportions of proteins in their secretomes are highly speci c.
Although no straightforwardly matching data are available from the literature, the obtained results can be indirectly compared with other studies. For instance, the obtained NPC secretory pro les show 71% overlap with the results of Mendes-Pinheiro et al. (2018), who applied proteomic approach to the primary cultures of human neuronal progenitors and identi ed 538 secreted proteins [43]. Signi cant overlaps of the obtained GPC secretory pro les with the corresponding data for primary cultures of human astrocytes should be noted as well [44][45][46][47][48][49].
Both NPC and GPC cultures secreted low amounts of neurotrophins (BDNF, CNTF, NGF). However, at the mRNA level, GPC cultures showed higher expression of BDNF, CNTF and NGF, which is consistent with higher rates of neurotrophin secretion by GPC cultures reported elsewhere [50][51][52].
GPC Thus, the results indicate pronounced neuroprotective, anti-in ammatory and pro-angiogenic properties of GPC secretomes as compared with NPC secretomes. The difference is apparently related to unique secretory pro les of the glial progenitor cells. One of the major causes of cell death in hypoxia is the accumulation of reactive oxygen species which results in oxidative stress [53,54]. According to the presented proteomic data, GPC secretomes show stronger involvement in the redox chemical processes than NPC secretomes. The difference in cytoprotective properties between the two cultures can be also due to higher representation of anti-apoptotic signal molecules in GPC-CM (e.g. LIF, HSPA4, HSPH1,  [57]. The therapeutic effects of CM were also evaluated in vivo. For instance, administration of NSC-conditioned media to the rats with experimental ischemic stroke promoted alleviation of neurological de cit and reduction in the infarct volume compared with the control group [58]. Intra-venous infusion of the bone marrow MSC-CM to the rats with experimental ischemic stroke signi cantly enhanced functional recovery and alleviated the microglial/macrophage in ltration of brain tissue [59]. As demonstrated by Hicks et al. (2013), the transplanted primary NSCs actively produce angiopoietin 1 (Ang1) which promotes an increase in the amount of microvessels in the ischemized area of the brain [60]. However, the effects of secretomes of the iPSC-derived neuronal and glial progenitors on the viability and structural organization of the SH-SY5Y cell cultures, as well as the recovery of the brain tissue after experimental ischemic stroke are reported here for the rst time.
This study also rstly demonstrates the effects of the hiPSC-derived GPC-CM and NPC-CM on the

Conclusion
Conventional therapies for vascular diseases of the brain do not allow complete functional recovery of damaged tissues, hence the modest success of neurological rehabilitation. Post-ischemic repair of brain tissue remains a primary challenge for modern medicine.
Secretory activity of the hiPSC-derived NPCs and GPCs was studied by enzyme-linked immunosorbent and proteomic approaches. The study reveals signi cant differences in representation of regulatory proteins (neurotrophins, growth factors, cytokines, heat shock proteins) in GPC and NPC secretomes. The

Consent for publication
Not applicable

Competing interests
The authors declare that they have no competing interests.

Availability of data and materials
All data analysed during this study are included in this published article with the exception of the data of proteomic analysis. Proteomic analysis data are presented in supplementary materials.

Funding
The study was supported by the Ministry of Science and Higher Education within the State assignment for the Research Centre for Medical Genetics.